NOTE: This is essentially the manuscript I sent Zenit, but with the Dutch quotations translated. Their extremely helpful editor, Eddy Echternach, provided the new title seen here; it involves a little play on words, as “between ship and shore” in Dutch is like saying “falling through the cracks” or “between two stools” in English, suggesting the intellectual homelessness of this interdisciplinary phenomenon as well as the physical location of its cause in the marine boundary layer. (The Dutch phrase “tusschen wal en schip” really means “to be lost inadvertently or through carelessness,” as it originally referred to the loss of cargo between ship and shore upon loading or unloading.)
I have made a few changes to make this version a little closer to the article printed in Zenit. As the Editor has inserted a comment in [square brackets], I have put a few explanatory comments of my own in [[double square brackets]].
The article as printed is very nicely laid out, with the various figures and the sidebar conveniently intermixed with the text. The Editor has very kindly made available his scans of the photographs for use here as well. However, the printed versions have considerably better resolution than the images shown here. The layout in the magazine is also much more attractive than here on the Web; but at least this makes the text available in English. Though particular pictures were not specifically mentioned in the text, I have tried to include them here in roughly the same places they appear in the magazine.
Bear in mind that this was written for a Dutch audience; consequently, there are extensive quotations from the Dutch GF literature, and references to Dutch writers on green flashes whose work has been particularly influential.
Thanks again to Eddy Echternach for his fine work in preparing the material for publication, and especially his skillful translation of my words into Dutch!
Beautiful mock-mirage flash, photographed by the author on 7 January 1996 from his “usual place” at Torrey Pines. (Torrey Pines is a 100-m high cliff on the southern California coast, not far from the university in San Diego.) [[NOTE: there are three universities in San Diego with similar names: San Diego State University, where I am an adjunct faculty member; the University of California at San Diego, close to Torrey Pines; and the University of San Diego, irrelevant here but still confusing!]]
The sudden color change is due to a combination of atmospheric scattering, which greatly weakens the blue and violet transmitted near the horizon; atmospheric absorption, mostly due to ozone, which removes most of the orange; and the dispersion of refraction, which makes the red image disappear before the green. At sunset, the bright setting Sun usually bleaches much of the red-sensitive photopigment from the retina, so that the image appears green somewhat sooner than it otherwise would; thus flashes that appear green to the eye often turn out yellow on color film.
I have been observing green flashes and studying their literature for several years, and have been impressed by the disproportionate influence of a few Dutch sailors and scientists. To my surprise, even some modern Dutch scientists are unaware of their remarkable heritage in this fascinating subject. To be sure, most people around the world first learn of green flashes from the account in Minnaert's wonderful book “Light and Color in the Open Air,” which has been translated into many languages. But Minnaert's account is based primarily on the work of two earlier Dutchmen, Marten Edsge Mulder and Pieter Feenstra Kuiper; and their work, in turn, depended on the studies of earlier Dutch workers.
This is a brief history of green-flash studies, together with a sketch of our present understanding of these beautiful and varied phenomena. The largest body of observations is that sent by Dutch seamen to Prof. A. A. Nijland; so let us begin with these.
The story begins with the Dutch expedition to Sumatra to observe the total solar eclipse of May 18, 1901. Among the members were an astronomer, Albert Antonie Nijland, and a physicist, W. H. Julius. On both the trip out to Sumatra and the return journey, they observed 3 green flashes at sunrise, and 4 at sunset. As one might expect from their training, the astronomer concluded that more observations were needed, while the physicist concocted a far-fetched theoretical explanation, which later proved to be worthless. We can forget about Prof. Julius's “anomalous-dispersion” theory, and concentrate on Prof. Nijland's remarkable next step.
Nijland realized, from his shipboard observations, that the best place to observe green flashes was at sea rather than on dry land. So he published an appeal to seamen in De Zee, saying that “No one is better placed to collect the data so badly needed about this phenomenon that is really still unknown and mysterious than a seaman on morning-watch or dog-watch.” He then provided detailed instructions for observers.
Many sailors responded to Nijland's appeal, but none so enthusiastically as Captain Egbertus E. Havinga, who turned in some 450 observations in the 3 years from Oct. 1902 to Oct. 1905, including almost 300 green flashes. Having received no response, and having decided “that there was not much more to discover,” he stopped recording his observations. Besides, as he explained, “A Steamship Line is no philanthropic institution; there is work to be done.” Still, in those three years, Captain Havinga contributed more green-flash observations than anyone before or since.
Nijland's appeal succeeded, but almost too well: he soon had more observations than he knew what to do with, and realized that this material would make a good thesis project for a graduate student. But apparently no astronomy student was interested in this half-meteorological phenomenon; and then the War intervened. Two decades elapsed before Nijland was able to find someone to analyze the great collection of observations he had collected.
Meanwhile, Marten Edsge Mulder, a professor of ophthalmology, had seen a green flash in 1907, but did not pursue his interest in the phenomenon until he read an account of a similar observation in the German magazine Kosmos in 1912. This prompted him to collect everything he could find about the phenomenon, and finally to publish it in 1922 as the first book ever devoted to this subject. In Hemel en Dampkring 19, 129-130 (1922), Mulder remarked that “The book is written in English because the English, as a seafaring nation, have made many observations and still show much interest in the phenomenon.”
Mulder's work is remarkable in several ways. Not only did he collect and organize the scattered accounts of green flashes; he was able to classify the observations into three groups, and this scheme has completely dominated all subsequent work on the subject. Most commendable of all, he was able to reject his initial impression that his own observation was an afterimage, and instead to realize that atmospheric dispersion is the basis of the phenomenon. Few devotees of the “afterimage” theory have been willing to give it up.
No doubt Mulder's division of green-flash phenomena into the green rim (which is always visible in a telescope, but never with the naked eye), the green ray that shoots up from the horizon, and the green segment, which he believed to be simply the green rim visible after the rest of the Sun has set, has had the most lasting effect on the subject. However, the latter category has turned out to be much more complex than Mulder and his followers imagined; it includes at least three separate phenomena, none of which is a simple segment cut off by the horizon.
The publication of Mulder's book prompted Havinga to complain that his work had been neglected, and stimulated Nijland to set a student to work on the collected observations. The student who finally did this was Pieter Feenstra Kuiper, whose thesis “De Groene Straal” [[literally, “The Green Ray”; but we would say “The Green Flash”]] added hundreds of observations to the available literature — though not a single one was made by him. [[However, see the letter from Cor Booy, who had a first-hand account of Feenstra Kuiper's green-flash observations.]]
In interpreting this rich material, Feenstra Kuiper relied heavily on Mulder's book, and on an independent and nearly simultaneous review of the literature by the American astronomer Willard J. Fisher (to whom I shall return presently.) In his thesis, Feenstra Kuiper published not only a summary of hundreds of green-flash observations, but an analysis of their durations, finding that the typical flash lasts just under 2 seconds. He also noticed that a few flashes — perhaps one in a hundred — last much longer, with several between 10 and 15 seconds.
Unfortunately, he also misunderstood a work on mirages by Alfred Wegener, the father of continental drift, and thus misinterpreted one common form of green flash as due to thermal inversions above the observer. Only recently have these been explained as due to inversions below eye level. He also adopted Mulder's somewhat misleading classification scheme. Finally, on the basis of rather flimsy evidence, he concluded “that a high humidity is in general favorable for the phenomenon.”
Feenstra Kuiper's thesis came to the attention of Prof. B. G. Escher, brother of the remarkable artist M. C. Escher. Prof. Escher was a geologist, so it was natural for him to plot the geographic distribution of the observations in Feenstra Kuiper's thesis on a map. He immediately saw “that the frequency of observations … is greatest in the South-Eastern part of the Mediterranean Sea, in the Suez Canal, in the Gulf of Suez, in the Red Sea and in the Gulf of Aden.” He therefore concluded that green flashes are most common “where the precipitation is the least, i.e. where the shipping routes pass near or between deserts,” and “that a dry climate is favorable for the appearance of this remarkable phenomenon.”
This remarkable conclusion is completely in accord with what we know today: that green flashes are usually by-products of mirages, which are favored by the large diurnal temperature changes in desert climates, where the low humidity allows strong radiative cooling at night. But it contradicted a major conclusion of Feenstra Kuiper.
Furthermore, Escher prefaced his own classification proposal, more detailed than Mulder's, with the modest apology, “A complete classification could only be given by someone who has made very many observations himself and moreover is well grounded in the natural sciences.” Feenstra Kuiper apparently took this remark as directed against himself personally, and angrily responded: “To observe astronomical and atmospheric phenomena does indeed demand much routine, and explanations of them, scientific preparation; but one would expect these sooner from an astronomer and meteorologist, than from a geologist.” He even went so far as to complain of “the authoritarian tone” of Escher's paper, which an impartial reader can hardly discern.
Perhaps if Minnaert had read this exchange, he might have given less weight to both Mulder's classification and Feenstra Kuiper's claims. But few astronomers read the proceedings of a Geographical Society. So Minnaert's account in “Light and Color in the Open Air” is mainly based on the works of Mulder and Feenstra Kuiper. And, in turn, most later writers have deferred to Minnaert; indeed, O'Connell (whose book is regarded by astronomers as the last word on the subject) did not even read Mulder's book. So the “textbook” account of green flashes today is largely based on these works.
In any case, it is fortunate that Minnaert chose to reproduce from Feenstra Kuiper's thesis the remarkable drawing made by D. P. Lagaaij, showing a “green ray” shooting up from a green flash. This unusual type of display, similar to a “green column” seen by Escher, is evidence that things are not so simple as the textbooks would have one believe.
The standard story, which goes back to David Winstanley's 1873 telescopic observations of the green rim, is that normal refraction separates the red and green images slightly at the horizon, so that when the red image is just below the horizon, the green rim remains visible. However, the green rim is only about 1/2 to 1/3 of a minute of arc wide, which projects to about half the width of a cone in the retina. [The cones are the light-sensitive cells with which we see colors — Ed.] Nevertheless, this “green rim” model of green flashes was popular among astronomers and physicists who had never seen an actual flash, such as the younger Lord Rayleigh, whose 1930 paper leans heavily on the 1906 account given by Arthur Rambaut.
As Rayleigh's paper was one of the few cited by Minnaert in his book, it has continued to be influential. So this model was still put forth by D. J. K. O'Connell in his 1957 book, “The Green Flash and Other Low Sun Phenomena,” although it hardly accounts for the color photographs shown there. (The phrase “low Sun phenomena,” by the way, was coined by W. J. Fisher.)
In 1955 Gerhard Dietze showed that, in fact, the green rim is too small and faint to be visible to the unaided eye when it alone is above the horizon. His paper is much less well known than it should be, because it was missed by O'Connell, who searched only the astronomical and not the meteorological literature.
So, if green flashes are not just a “green segment” cut off by the horizon, what are they? The first clue is buried in the earliest known scientific paper on the subject, a note sent by James Prescott Joule to the Manchester Literary and Philosophical Society in 1869. Joule casually remarked that “at the moment of the departure of the sun below the horizon, the last glimpse is coloured bluish green.” But he also provided a sketch of the setting Sun with little “feet” where it joined the horizon — a shape likened to the Greek capital letter Omega by the Italian astronomer Riccò much later.
These “feet” are nothing but a reflected image of the part of the Sun's disk immediately above them. This reflection is produced by the common inferior mirage, familiar to everyone today because of its visibility on asphalt roads on sunny days. This is the mirage of the desert, and of the sea as well; indeed, the term “mirage” was used by French sailors long before it entered the technical literature, even in French. Whenever the water is a degree or more above the temperature of the air, the mirage can be seen at sea.
Unfortunately, Joule's note is so little known that even O'Connell was unaware that it was ever published. It remained for the English astronomer John Evershed to re-discover the connection between the inferior mirage and Joule's “last glimpse” green flash. Though he first mentioned the word “mirage” in connection with a green flash in 1915, it was not until 1923 that he began to insist that the two phenomena are related. “It seems evident,” he wrote, “that the mirage layer greatly intensifies the ordinary dispersion effect, by adding the light from the reflected image to the direct image at the moment of setting. The normal dispersion effect at sunset under conditions when there is no mirage is scarcely visible to unaided vision, although easily seen in a telescope of low power.”
Independently, an American astronomer, Willard J. Fisher, was also connecting mirages with green flashes. It was he who coined the phrase “low Sun phenomena” that O'Connell took for the subtitle of his book. Fisher first became interested in sunset phenomena in the Philippines, where he noticed that the apparent duration of sunset is sometimes much longer than predicted from Nautical Almanac data. While carefully observing sunsets to confirm this fact, he noticed that there are two distinctly different kinds of marine sunset:
“Type A occurs always when mirage is perceptible … As the descending sun, vertically compressed by atmospheric refraction, approaches the sea horizon, a protuberance … grows out below, and almost simultaneously a line of light appears in the sea horizon, which lengthens horizontally and thickens upward till the protuberance and the thickened line join … For about half the duration of sunset the sun presents the appearance of an inverted fish globe, whose mouth widens to the sun's diameter; from then on the vanishing disk looks like an ellipse much flattened below, and vanishes as a small oval spot… This spot does not sink below the horizon; it `goes out,' above it.”
This inferior-mirage flash was seen on 30 March, 1998, at Scripps Pier (Southern California), a place that lies about eleven meters above sea level. Notice the (disappearing) yellow spot in the middle of the green flash. This is the kind of green flash that can be seen from the Low Countries. [[NOTE: this image is highly magnified. For a comparison with an earlier stage in this same sunset that shows the full disk of the Sun, click here. I have also added an animation of such a sunset since this paper was written; and, more recently, simulated images with more realistic colors.]]
“Type B never occurs simultaneously with a perceptible mirage. The descending sun flattens below as it approaches the horizon, which is not easy to see under it, being comparatively dark. The corners of the disk as it passes below the horizon are rounded, instead of projecting, until about midsunset; then the remaining segment begins to show a rim, so that it looks for a time not unlike the `tin hat' or trench helmet of the American Expeditionary Force. This flattens down in the middle faster than it shortens horizontally, becomes a line of light and disappears in dots and dashes among the waves, if the horizon is near. The disappearance is slow, not like the vanishing of the spot at the end of type A.”
Fisher soon found that green flashes were common in type A sunsets and rare in type B. However, he seems not to have understood the mirage effect in detail. What happens is that the inverted image of the mirage blends smoothly into the erect image above it; at the line where they join, there is infinite vertical magnification. The large vertical stretching of the image near this line expands the normally invisibly thin green rim into a fat blob, often several minutes of arc in vertical span. The lower part of this blob is, of course, the inverted image of the upper part, reflected in the mirage. [[In fact, if one assumes the temperature profile prescribed by the Monin-Obukhov similarity theory of boundary-layer meteorology, all the features described by Fisher (and many others) are beautifully reproduced.]]
Though Fisher was aware of the connection between mirages and warm surfaces, he did not explicitly connect his type A and B sunsets with the difference in temperature between water and air. However, Captain Havinga did. He independently had noticed the difference between the two types of sunsets, and explicitly stated that “… when the temp. of the air is lower than that of the water, and the Sun is a few min. [[of arc]] above the horizon, a counter-Sun rises; the Sun sinks, the counter-Sun rises, they fuse together and the Sun takes on the so-called balloon-form. But the last little segment disappears not with sharp corners on the horizon, but as a little ellipse a few min. above the horizon.” Like Fisher and Evershed, he recognized that this type of sunset favors green flashes.
The computer simulations depicted here give an overview of the course of the most important green-flash phenomena. From top to bottom are the inferior-mirage flash, the mock-mirage flash, and the sub-duct flash. The figures show only the (mostly overlapping one another) contours of the Sun for wavelengths of 530 nm (“green”) and 700 nm (“red”); the astronomical horizon is marked by a dashed line, and the apparent horizon by a solid line. [[The scales at the sides are in arc minutes; the number below each sub-panel is the true altitude of the Sun's center at the time for which the outlines are shown.]]
The [[transfer-curve]] graphics following the simulations illustrate how light of these two wavelengths are refracted differently by the atmosphere; the true altitude above the horizon is plotted against the apparent altitude. The divisions on the axes are in arc minutes. [[On p.252 of my Zenit article, the three kinds of flash are all shown together for comparison, with the transfer curve for each flash beside its simulation; but the limited resolution of computer screens makes that impractical here. So I show the inferior-mirage flash and its transfer curve above, and place the other two flashes in the text below, closer to where they are discussed. See also the sidebar from the article.]]
If the green flash produced by the inferior mirage were the only kind of green flash, the subject would have been much less confused. There are actually 5 or 6 kinds of green flash, however. The second most common one was described in the last sentence of Joule's letter: “Just at the upper edge, where bands of the sun's disk are separated one after the other by refraction, each band becomes coloured blue just before it vanishes.”
This type of flash accompanies a recently-understood phenomenon I have called a “mock mirage.” Unlike the classical inferior and superior mirages, which can crudely be thought of as caused by internal reflection at a less-dense layer of air, the mock mirage is a purely refractive phenomenon. When a thermal inversion lies below eye level, the denser air below it acts like a big, weak lens. Near grazing incidence, this lens has enough power to form an inverted image of a strip of the sky beyond it. This inverted strip of sky can produce a double image of the limb of the Sun. When the bottom edge of the Sun reaches this double-image almucantar, an elliptical red flash appears below the main solar image; when the upper limb reaches it, a sharp-cornered strip appears to separate from the limb and turn green before vanishing.
These flashes can be much longer-lasting than the inferior-mirage flash. There are reliably timed observations as long as 15 seconds; I myself have seen a 5-second flash. However, as they require appreciable thermal structure below eye level, they are rarely seen near sea level; most observations are from cliffs or seaside hills at least some tens of meters high. Consequently, most of the phenomena in O'Connell's book, which were photographed from Castel Gandolfo, 450 meters above the sea, are of this type. They are also commonly seen from mountain observatories in Mediterranean climates, such as those on the western coasts of the Americas.
In contrast, the inferior-mirage flash is best seen from heights on the order of 3 to 5 meters, and gradually become more compressed and less conspicuous at considerable heights. Thus, the deck of a ship is nearly ideal for observing inferior-mirage flashes; so these dominate the list in Feenstra Kuiper's thesis. Hence, most of the flashes seen in the Low Countries correspond to Joule's “last glimpse” of the setting Sun.
There are still other varieties of flash, most of them associated with optical ducts, which we have no room for here. But Mulder's “green ray” type of display deserves a few words, as these are occasionally seen from near sea level. There are several very reliable accounts of these phenomena, which I believe are crepuscular rays in hazy air illuminated by an unusually large green flash. The ray is formed by forward scattering by the haze particles; if it passes just above the observer's head, it can appear to “shoot up” from the horizon, as described by Mulder, Escher, and others.
The sub-duct flash:
Considering the complexity of green flashes, with several different forms, each having its own distinct optical mechanism, and the color of each being produced by a combination of refraction, dispersion, scattering, absorption, and (at sunset) the partial bleaching of photosensitive pigments in the retina, it is no wonder that our understanding of these phenomena remains incomplete. Because of the complexity of forms and mechanisms, understanding is possible only when all the various types are treated separately; and, as Escher pointed out, this classification is only possible when one has seen many different green flashes.
Alas, no one person can see all the variety made possible by different geographical settings and climates. So I have found it absolutely essential to read as many first-hand accounts of observations as possible. Captain Havinga, probably the all-time champion observer of green flashes, once complained that the trouble with most accounts is that “All the letter-writers have heard of the green flash now and again, don't know what and how much has already been written about this subject, have seen the phenomenon a single time and suppose that they have made an important discovery worth reporting.”
It is ironic that if everyone had taken this remark to heart, we would not have the great body of independent observations that exist today. Although observations from experienced observers are much more useful than those of a first-timer, it is also true that most of the rare forms of flash are reported by first-time observers (like Mulder, with his green-ray observation). Such remarkable and well-established phenomena as green-ray displays, and flashes lasting over 5 seconds, which make up only about 1 per cent of all green-flash reports, would not be known if everyone took a conservative attitude toward publication.
The complexity of the subject makes simplistic explanations, such as those found in the textbooks, quite inadequate. The slow understanding of green flashes is thus partly due to the difficulty of assimilating all the different disciplines that are involved: astronomy, meteorology, hydrography, atmospheric optics, physiological optics, and others. As Cornelis said of the related problem of great refraction at high latitudes, “Through the division among all sorts of professions and specialties the refraction has been lost in the crowd. The meteorologists leave the teaching to the astronomers, but these have basically an abhorrence of everything close to the Earth, and are concerned with refraction only when it has an insignificant value. Also the meteorologists are hardly attracted by the so-called terrestrial refraction, because it is parcelled out by the compartmentalization to geodesy, while the geodesists have not extended their activities to the polar seas.” So it is with green flashes; no field wants to “own” them, so they have fallen through the cracks. [[This was the phrase that Eddy Echternach ingeniously translated as “tusschen wal en schip” — literally, “between shore and ship”. So his version of my original sentence says literally: “So it is with the green flashes: they belong nowhere and have literally fallen between shore and ship.”]]
Further progress depends on more and better observations today, just as it did when Prof. Nijland appealed for observations. Visual observations by experienced observers are still useful; however, binoculars should be used, as most flashes are too small to be seen by the naked eye. Today, color photography makes possible permanent, accurate records of sunsets. Green flashes are usually small, so a long telephoto lens should be used. We still need photographs of green-ray displays and other unusual flash phenomena, such as the very large and/or long-lasting flashes that seldom occur. An educated amateur can still contribute to scientific research in this field, so I encourage everyone to become familiar with the common green flashes, and to watch for what Mulder called the “rare and abnormal forms” that have yet to be understood.
The author thanks the Atmospheric Sciences Division of the U. S. National Science Foundation for their financial support of the green-flash research. [[Dutch]] translation: Eddy Echternach
A. A. Nijland, “Over den `groenen straal' en eenige andere hemelverschijnselen,” De Zee 24, 60-68 (1902)
E. Havinga, “De groene straal,” Hemel en Dampkring 19, 161-165 (1922)
Prof. Dr. M. E. Mulder, “De groene straal en de groene branding,” Hemel en Dampkring 19, 129-130 (1922)
Prof. Dr. M. E. Mulder, “The “Green Ray” or “Green Flash” (Rayon Vert) at Rising and Setting of the Sun,” (T. Fisher Unwin, Ltd., London, 1922)
Pieter Feenstra Kuiper, “De Groene Straal” (C. de Boer Jr., Helder, 1926)
B. G. Escher, “De Groene Straal,” Tijdschr. Kon. Ned. Aardr. Gen. (2) 47, 602-609 (1930)
P. Feenstra Kuiper, “De Groene Straal,” Tijdschr. Kon. Ned. Aardr. Gen. (2) 47,868-870 (1930)
M. Minnaert, “The Nature of Light and Color in the Open Air,” (Dover, New York, 1954) [ EDITORS: You will probably want to cite the Dutch edition: “De natuurkunde van't vrije veld door M.Minnaert” (Thieme,Zutphen, 1937) — or perhaps a later one. ]
David Winstanley, “Atmospheric Refraction and the last rays of the Setting Sun,” Proc. Manchester Lit. Phil. Soc. 13, 1-4 (1873)
Lord Rayleigh, “Normal atmospheric dispersion as the cause of the "Green Flash" at sunset, with illustrative experiments Proc. Roy. Soc. A 126, 311-318 (1930)
A. A. Rambaut, “The green flash on the horizon,” Symons's Met. Mag. 41, 21-23, 41-45 (1906)
D. J. K. O'Connell, “The Green Flash and Other Low Sun Phenomena,” (North Holland, Amsterdam, 1958)
G. Dietze, “Die Sichtbarkeit des `grünen Strahls',” Zeitschr. f. Meteorologie 9, 169-178 (1955)
J. P. Joule, “On an appearance of the setting sun,” Proc. Manchester Lit. Phil. Soc. 9, 1 (1869)
J. Evershed, “The green flash at sunset,” Nature 111, 13 (1923)
W. J. Fisher, “Low-Sun phenomena in Luzon III. Marine sunsets and the duration of sunset on Manila Bay and the China Sea,” Philippine J. Sci. 17, 607-614 (1920)
W. J. Fisher, “Low-Sun phenomena — IV. The `Green Flash' ” Pop. Astron. 29, 251-265, 382-392 (1921)
A. T. Young, G. W. Kattawar, P. Parviainen, “Sunset Science. I. The Mock Mirage,” Appl. Opt. 36, 2689-2700 (1997)
E. Havinga, “Groene straal en kimduiking,” Hemel en Dampkring 32, 114-118 (1934)
W. Cornelis, “Het Nova-Zembla-verschijnsel,” De Zee 46, 194-201 (1924)
You might also like to read the translation of a letter from Cor Booy that was published in response to this article. He tells a wonderful story about Feenstra Kuiper, who gave lectures in the 1950s that Mr. Booy attended.
© 1999, 2006, 2014 Andrew T. Young